[Rivet] Analyses extended with more data

Andy Buckley andy.buckley at durham.ac.uk
Sun Jul 19 23:44:01 BST 2009


Frank Siegert wrote:
> Hi Riveters,
> 
> What is our policy regarding analyses that have been updated later with 
> more data, do we want them in Rivet at all, regardless of how similar 
> their method/result is to the updated one? One such example is 
> 
> http://arxiv.org/abs/hep-ex/0512020 (CDF_2006_S6450792)
> http://arxiv.org/abs/0807.2204 (CDF_2008_S7828950)
> 
> (though I haven't checked the details of their midpoint cone algorithm for 
> differences)
> 
> The latter is already in Rivet, and the former I have ready to commit, but 
> don't know whether I should. If it should go in (i.e. if I don't hear any 
> contrary opinions), I'm definitely going to put a notice about the update 
> into the analysis description. 

That seems the most useful approach to me: when choosing analyses to 
write, we should mainly prioritise those for which there is no current 
~equivalent, but I'm very happy to see updates go in as well, 
particularly if superceded analyses are noted as such. When used in 
tunings, of course, we should be aware that analyses are not 
independent, and down-weight accordingly in calculating the chi2. But I 
can see no reason to not be comprehensive if the opportunity presents 
itself!

Andy

-- 
Dr Andy Buckley
Institute for Particle Physics Phenomenology
Durham University
0191 3343798 | 0191 3732613 | www.insectnation.org


More information about the Rivet mailing list